Article

Psychology of Music
41(1) 75–88
Free improvisation and © The Author(s) 2011
Reprints and permission: sagepub.
performance anxiety co.uk/journalsPermissions.nav
DOI: 10.1177/0305735611415750
among piano students pom.sagepub.com

Robert Allen
Music Director, Bob Allen Music, Folsom CA, USA

Abstract
The purpose of this study was to compare the levels of anxiety that students experienced according
to whether their public performance consisted of a free improvisation or a repertory piece. The
researcher had two objectives: (1) examine the relationship of students’ levels of anxiety to free
improvisation and repertory pieces during a performance; and (2) examine the effectiveness of free
improvisation as a treatment for the reduction of performance anxiety. The following instruments
were used for data collection: (1) Spielberger’s (1973) State-Trait Anxiety Inventory for Children; (2)
Musical Anxiety Report Scale; (3) subject interviews; (4) parent questionnaire; and (5) performance
video. Thirty-six elementary, middle and high school students were selected from a list of potential
subjects. Participants were selected based on age (7–18 years) and years of training (1–8), as well
as to comprise a gender balance. Sample criteria required that all subjects who participated in the
study: (1) played the piano; (2) claimed to have experienced music performance anxiety; and (3) had
not received any previous psychological or pharmacological treatment for their music performance
anxiety. Results from this study validated free improvisation as a treatment for significantly reducing
anxiety during the public performance of a musical work.

Keywords
free improvisation, gender, improvisation, Musical Anxiety Report Scale (MARS), performance anxiety,
State-Trait Anxiety Inventory for Children (STAIC)

Introduction
Most musicians have experienced the sensations of a dry mouth, knot in one’s stomach, lump in
one’s throat, that tingling feeling referred to as the “butterflies,” or perspiration in the palms of
their hands. These feelings are generally referred to as symptoms of “performance anxiety.”
These sensations distract the performer and can negatively affect the outcome of a musical
performance: (1) a dry mouth may cause a singer to begin a phrase with a raspy tone; (2) a knot

Corresponding author:
Robert Allen, PO Box 1372, Folsom, California, 95763 USA.
[email: robertallenjr@yahoo.com]

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

76 Psychology of Music 41(1)

in one’s stomach may cause a trombone player to produce a weak tone; or (3) perspiration in the
palms of one’s hands may cause a drummer to lose his grip on his sticks while playing (Allen,
1996). These examples are the outward manifestations of inner feelings generally referred to as
“performance anxiety,” which is a condition that can be so severe as to cause students to
discontinue musical studies.
When musicians perform in public, psychological factors come into play (Ross, 2007).
During performances of a preconceived musical product whose interpretation is predetermined,
levels of anxiety seem to increase (Sawyer, 2000). The performer’s perception of what the
listener expects may be at the center of performance anxiety. In a previous study, Roland
(1993) found that performers experienced significant distress and marked impairment when
performing in front of an audience. Rack (1995) identified specific symptoms of anxiety as
debilitative: self-doubt, all-or-nothing thinking, and obsessive worrying. These particular
symptoms weakened subjects’ ability to perform. In addition, Kim (2003) described stress and
tension as two main characteristics of anxiety.
Lehrer (1987, p. 143) stated that “stage fright” seems to comprise “fear of fear, fear of social
disapproval, problems with distraction, and a judgmental attitude.” Furthermore, there seemed
to be considerable evidence that stage fright is multifaceted; “in addition to partially independent
somatic, cognitive, and behavioral manifestations, stage fright seems to comprise fear of fear,
fear of social disapproval, problems with distraction, and a judgmental attitude” (p. 144). For
this study, performance anxiety was defined as a performer’s state of being uneasy, apprehensive,
or worried about what may happen in a musical performance and the physical or physiological
consequences of that anxiety. It is believed that an improved understanding of performance
anxiety, if coupled with empirical research on music learning processes (i.e., Gordon, 1997;
Campbell, 2009), might aid in the development of an effective approach for overcoming anxiety
associated with playing music in the context of public performance.
Researchers have examined specific aspects of intervention, such as psychological,
pharmacological, and physiological treatments, as means for reducing levels of anxiety that
musicians experience in relation to musical performance. The published literature related to
anxiety offered evidence of the aforementioned treatments as means for reducing levels of
anxiety during a musical performance. Quantitative studies of performance anxiety indicate
a continued trend away from psychotherapy that began in the 1970s. Other studies showed
that the use of beta-blockers reduced performance anxiety, particularly among string and
keyboard players (Steptoe, 1982). Appel (1976) compared systematic desensitization with
musical analysis and performance rehearsal and a no-training control group. Following the
treatment, the systematic desensitization group performed at lower subjective anxiety levels,
exhibited lower pulse rate, and made fewer performance errors than the other groups.
Lehrer (1987) concluded that the various forms of performance anxiety are treatable by a
specific intervention. For example, relaxation therapies and drug therapy affect somatic symptoms.
It is unclear to what extent, if any, that drug therapies may have the potential for side effects,
including some that are deleterious to performance. Although studies of such interventions do
exist, there is little to prove or disprove the long-term effects of medication. Furthermore, Lehrer
found that there is little knowledge of “interactions between medication and various psychological
approaches to managing somatic manifestations of stage fright” (p. 145). Although behavioral
interventions in conjunction with an increase in the frequency of performance experience may
have a beneficial effect on stage fright, little empirical evidence exists to support this view.
Previous research in the areas of anxiety reduction regarding musical performance
concludes that anxiety should be reduced in order for a musician to perform at his or her
optimum level (Hamann, 1985). LeBlanc, Jin, Obert, and Siivola (1997) tested 27 male and

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 77

female high school band members performing solos under three levels of audience presence:
(1) alone in a practice room; (2) in a practice room with one researcher present; and (3) tape
recorded performances in the rehearsal room with all researchers and a peer group. Results
found that, with each succeeding performance, self-reported anxiety rose and that each
reported increase was significant. Gender emerged as a significant predictor of heart rate
during performance. Female performers attained higher heart rates.
Ryan (2000) studied the differential responses of 26 sixth grade piano male and female
piano students to musical performance anxiety, measuring the heart rates, behavior,
performance quality, and anxiety levels before and during a piano recital. These variables were
examined for possible gender differences. Results found that several relationships that emerged
in analysis were driven by a strong relationship in a single gender, but were nearly absent, or in
fact opposite, in the other.
Wilson and Roland (2002) suggested that hypnotherapy and the Alexander Technique may be
effective in reducing performance anxiety. Results showed that more than half of performers
experience significant distress and marked impairment when performing in front of an audience.
Studies comparing degrees of performance anxiety between sexes found that there is equal
incidence among males and females (Hamann, 1982; Ryan, 2000). Wilson (2002) proposed three
independent sources of stress: (1) trait anxiety – any personality characteristics, constitutional or
learned, that mediate susceptibility to stress; (2) situational stress – environmental pressures such
as public performance, audition, or competition; and (3) task mastery – ranging from performances
of simple, well-rehearsed works to those of complex, underprepared material.
Recent studies have attempted to evaluate the use of treatments based on free improvisation
in order to reduce the extent to which an audience affects levels of performance anxiety. Kim
(2003) examined the effects of two music therapy approaches on ameliorating the symptoms
of performance anxiety: (1) improvisation-assisted desensitization; and (2) music-assisted
progressive muscle relaxation and imagery. Thirty female college students were randomly
assigned to one of two groups (N1 = 15, N2 = 15). The improvisation-assisted desensitization
treatment was applied to group 1 and the music-assisted progressive muscle relaxation and
imagery treatment was used for group 2. Four types of visual analogue scales were used: Music
Performance Anxiety (MPA), stress, tension, and comfort. When the result of each condition
was compared from pretest to posttest, Kim found that there were six statistically-significant
results out of seven measures. Although all of the subjects from group 1 were able to successfully
complete the study, several had difficulty improvising. Both treatments provided significant
reductions in performance anxiety among participants, but neither treatment was significantly
more effective than the other. Although the aforementioned treatments have had encouraging
results, little has been documented in regard to free improvisation as a treatment.
During the 1990s, a trend towards the use of music itself to reduce anxiety began to develop
based on free improvisation methods (Darling, 2008). Free improvisation emphasizes the cre-
ative process, putting the decisions concerning musical content at the discretion of the per-
former, reducing or eliminating predetermined expectations, thus lowering levels of
performance anxiety (Furth, 1969). For this study, free improvisation was defined as

a spontaneous musical creation produced in an unthinking state, one in which we are relaxed yet
aware, incorporating and negotiating disparate personal perspectives and worldviews not limited by
genre or methodology, applied to a wide range of highly personal, individual styles. (Allen, 2010, p. 41)

The purpose of this study was to compare the levels of anxiety that students experience accord-
ing to whether their public performance consisted of a free improvisation or a repertory piece.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

78 Psychology of Music 41(1)

This was based on the assumption that a free improvisation instructional strategy might aid in
reducing levels of performance anxiety among school-aged piano students, and should there-
fore be systematically evaluated for its potential to offer improvements to music education. The
researcher examined the extent to which specific criteria based on free improvisation affected
levels of anxiety when students performed without an audience. Rather than gathering physi-
ological data that might be correlated to performance anxiety, self-report tests such as
Spielberger’s (1973) State-Trait Anxiety Inventory for Children and the Musical Anxiety
Report Scale were used to provide purely-subjective data based on perceptual experiences.
The primary benefit of this study is that it may contribute to a better understanding
of performance anxiety among musicians (Allen, 2010). Although some musicians have
developed strategies for applying free improvisation to performance anxiety (Darling, 2008),
this study appears to be the first study to test and measure the applicability of free improvisation
as an individual treatment to problems of performance anxiety. This study had the following
objectives: (1) examine the relationship of students’ levels of anxiety to free improvisation and
repertory pieces during a public performance; and (2) determine the extent to which instruction
in free improvisation is an effective treatment in reducing performance anxiety.
In this study, in which elementary, middle, and high school piano students were taught a set
of specific musical skills, there were two independent variables. The first independent variable
was the treatment in which subjects developed a free improvisation using a specific skill set of
musical elements. The second independent variable was the control condition in which a
control group did not receive the treatment. The aforementioned piano students were randomly
assigned to perform a free improvisation, free improvisation and/or repertory piece, or a
repertory piece only, in public. The purpose of the experiment was to determine the effect that
free improvisation had on a single dependent variable: state anxiety. In addition, this study
employed two types of dependent measures: (1) the State-Trait Anxiety Inventory for Children
(STAIC) – used to measure participants’ state and trait anxiety related to music performance
situations; and (2) the Musical Anxiety Report Scale (MARS; Allen, 2010) – designed by the
researcher as a way to measure each performer’s personal views regarding the effects of free
improvisation on their levels of anxiety with or without an audience. In addition, student
interviews, conducted by the researcher, supported the findings from the STAIC and the MARS.

Method
The design of this study was mixed methods with an emphasis on the quantitative element.
Although standardization and objectivity allow for statistical inferences, the inclusion of visual
and aural observations provides additional data from which individual cases may be examined
for uniqueness. While the State-Trait Anxiety Inventory for Children (STAIC) and Musical
Anxiety Report Scale (MARS) were used to measure the effects of free improvisation on
performance anxiety, the qualitative component enabled systematic consideration of how each
subject’s feelings towards free improvisation and anxiety were influenced to some degree by the
social, psychological, and cultural environment in which they lived.
This study employed a “three-group randomized groups design” (Orcher, 2005, p.183): (1)
treatment group E1 consisted of subjects who received free improvisation instruction only; (2)
subjects from treatment group E2 received both free improvisation and traditional instruction;
and (3) control group E3 subjects received traditional instruction only. The use of simple random
sampling made the design “equivalent.” The design of this study was a “true experiment”
(2005, p. 183).

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 79

Thirty-six elementary, middle and high school students were chosen from a list of approximately
250 potential subjects who had no previous free improvisation training. The list consisted of past
and current students of the researcher. In order to eliminate “teacher effect,” all participants were
selected from the same studio. Sample criteria required that subjects: (1) played the piano; (2)
claimed to have experienced music performance anxiety; and (3) had not received any previous
psychological or pharmacological treatment for their music performance anxiety. All of the
participants had previous experience performing musical pieces on the piano in front of an
audience during recitals. This resulted in a gender balance of 50% boys and 50% girls. The age
range for elementary subjects was 7–11 years, for middle school subjects 12–14 years, and for
high school subjects 15–18 years. The elementary school subjects had studied 1–3 years, middle
school subjects 2–5 years, and high school subjects 2–8 years. They were randomly divided into
three groups (see Table 1).
Subjects from treatment group E1 were given the treatment, a skill set consisting of melodic,
harmonic, and rhythmic elements, combined with a limited formal structure based on a
series of five events. They developed a free improvisation during the individual sessions,
administered, observed and recorded by the researcher, once a week for six weeks. They
played a free improvisation during the performance. Treatment group E2 participants also
received the treatment. In addition, this group played two pieces each during the performance;
the first, a free improvisation followed by a standard repertory piece. Control group E3 did not
receive the treatment. Instead, they received traditional piano lessons and performed only a
standard repertory piece. Members of the control group were given the option of memorizing
their repertory piece.
The researcher utilized six kinds of data and one treatment. Quantitative instruments consisted
of the State-Trait Anxiety Inventory for Children (STAIC) and Musical Anxiety Report Scale
(MARS). MARS was first used in a pilot study conducted by the researcher (Allen, 2007). Subjects
were given 10 statements to consider, nine of which consisted of a choice of four possible
responses: 4 = strongly agree, 3 = agree, 2 = disagree = 2, and 1 = strongly disagree. They were
to choose only one answer. Qualitative components consisted of the parent questionnaire, subject
interviews, performance video, audio recordings, observations, and the treatment.
Twenty-four subjects from treatment groups E1 and E2 were given six individual sessions
during scheduled times, one per week for six weeks. The treatment was based on a set of specific
skills (skill set) related to free improvisation. Participants developed different aspects of free
improvisation during each session: session A – developing short melodies; session B – creating
tonal sonorities; session C – subjects played percussive-like hand-to-hand rhythms; session D –
form was explained as “events”; session E – subjects performed the “climax” during free
improvisations by utilizing figures, motives, rhythms, and registers; session F – review of the

Table 1.  Research subjects

Treatment Group E1 Treatment Group E2

(A) 2 elementary, 2 middle school, (A) 2 elementary, 2 middle school,

and 2 high school girls and 2 high school girls
(B) 2 elementary, 2 middle school, (B) 2 elementary, 2 middle school,
and 2 high school boys and 2 high school boys
Control Group E3
(A) 2 elementary, 2 middle school, and 2 high school girls
(B) 2 elementary, 2 middle school, and 2 high school boys

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

80 Psychology of Music 41(1)

skill set and performances of free improvisations followed by final comments. The STAIC
T-anxiety test was given to each subject during session E.
The following lesson plan for session A is representative of guidelines followed in other
sessions:
Treatment (Session A)

1. Introduction: This session consisted of a brief overview of concepts relating to free

improvisation and the skill set followed by a discussion concerning the subject’s musical
background and personal experiences with performance anxiety or stress.
2. Demonstration of basic design of events: (1) melody – short melody using scale material in
any combination of white or black notes; (2) harmony – any combination of two or three
note sonorities, dissonance resolving to consonance; (3) rhythm – small rhythmic group-
ings of notes, rhythms not to be repeated more than three times consecutively; (4) climax
consisting of the use of melodic, harmonic, and rhythmic elements from previous events;
(5) restatement – revisiting prior material from events 1 and 2 which may include new
material. The form of each free improvisation consisted of five sections, or events, based on
a sequence of register changes accompanied by parallel changes in harmonic texture,
rhythmic intensity, and dynamic contrasts. Registers and dynamics were used in order to
provide points of delineation, visually and aurally, for developing the overall structure.
3. Develop melody and dynamics relating to the first event. Free improvisation by the sub-
ject, which included improvising with their eyes closed, relying on audiation (hearing) to
create music.
4. Free improvisation: The researcher left the room, all the while taking note of the subjects’
use of skill set concepts while the subject played. After improvising at the keyboard, each
subject was asked to express his or her views about the experience.
5. Assignment and closing conversation: The researcher reviewed the session and directed
the subject to improvise several different melodies without regard to any particular
formula.

Results
Before the performance began, subjects took the STAIC State Anxiety pretest (the preperformance
score). After the pretest, each subject performed a free improvisation and/or repertory piece.
Immediately following individual performances, each subject then took the STAIC State Anxiety
posttest (the post-performance score).
Means distributions (Table 2) found that subjects with medium and high levels of T-anxiety
who performed a repertory piece (only E3) had higher levels of anxiety than subjects who were
given the free improvisation treatment (E1 and E2). On the other hand, control group subjects
(E3) who were categorized as having low T-anxiety exhibited total means lower than the
treatment groups. Subjects with high degrees of T-anxiety who performed a free improvisation
exhibited lower levels of anxiety than those who performed repertory pieces. Differences
between subjects with medium T-anxiety and those with low T-anxiety demonstrated that free
improvisation had a minimal effect on subjects who experienced lower levels of T-anxiety.
Although T-anxiety means distributions showed changes in levels of anxiety, results from four
analysis of variance (ANOVA) tests for subjects categorized as having high, medium, and low
anxiety levels between test groups E1, E2, and E3 were not significant (p > .05).
Single-tailed t-tests were used to determine differences in levels of anxiety before and after a
public performance of free improvisations and repertory pieces for all subjects. STAIC S-anxiety

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 81

Table 2.  Means for STAIC T-anxiety scale

Anxiety category Free Free Rep Row means

improv. improv./rep only
High 37.67 37.25 42.00 38.97
Medium 34.20 33.00 34.33 33.88
Low 29.00 28.67 27.00 28.22
Column means 33.62 32.97 34.44 N/A

Male Female
High 6 4
Medium 7 6
Low 5 8

pretest and posttest results found that subjects from both treatment groups had significant
(p < .05) reductions in state anxiety following the performance, while results were not significant
(p > .05) for the control group subjects. A regression analysis was used in order to determine the
strength of a correlation between the number of years that subjects studied and the change in
S-anxiety pretest and posttest scores. Subjects from treatment group E1 had a range of 1–7 years
of study. Subjects from treatment group E2 had a range of 1–5 years of study. Subjects from
control group E3 had a range of 1–8 years of study. Results found that all subjects experienced
some variation in individual levels of anxiety from S-anxiety pretest to posttest regardless of
their years of study. Regression analysis found no significant correlation between the number of
years that subjects studied and the change in S-anxiety pretest and posttest scores.
When subjects were divided by gender into high, medium, and low categories of T-anxiety, it
was found that more male subjects than female subjects were categorized as medium and high.
Forty-five percent of females had low levels of anxiety compared to 28% of males. Means for
female subjects from control group E3 were highest. T-anxiety levels of female subjects who
performed a repertory piece (E3) were higher than all other groups. T-anxiety scores between
male and female subjects from treatment and control groups were not significant. ANOVA
between males from both treatment groups and males from the control group found that
reductions in levels of S-anxiety were significant (p < .05), but not significant between females
from pretest to posttest. In addition, results from paired t-tests for paired observations of male
and female T-anxiety scores, male and female S-anxiety pretest scores, and male and female
S-anxiety posttest scores were not significant (p > .05). Overall, results confirmed that gender
had little effect on levels of anxiety during a musical performance in front of an audience.
Reductions in STAIC S-anxiety from pretest to posttest had a very weak association to gender.
Immediately following their performance subjects responded to a number of statements
regarding their feelings towards performing a free improvisation and/or repertory piece while
performing in front of an audience. In addition, they were asked whether they felt that the free
improvisation skill set helped them and others as well. Results of the Musical Anxiety Report
Scale (MARS) are shown in Table 3. The highest possible score was 36, the lowest, 9. Subjects
who scored 29.5 or higher regarding statements 2, 3, 4, 5, and 7 preferred improvising alone
and agreed that the skill set helped them and would be beneficial to others. Subjects who scored
in the 25–29 ranges had varied feelings about anxiety regarding statements 2, 3, 4, 5, and 7,
but preferred improvising alone. They also felt that the skill set was beneficial to them as well as
others. Subjects with scores in the range of 19–24.5 generally felt low anxiety regarding
statements 2, 3, 4, 5, and 7, but felt that the skill set was helpful to others as well.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

82 Psychology of Music 41(1)

Table 3.  Subject responses to MARS

Statement Number of responses

Strongly disagree Disagree Agree Strongly agree

  2. I have a fear of performing in front of 2 15 18  1
other people
  3. I have a fear of making mistakes 3  8 20  5
  4. I have a fear of forgetting the music 3 11 18  4
  5. P erforming a repertory piece in front of 5 15 12  4
people makes me nervous
  6. I mprovising makes me feel better about 2  8 15  7
performing in front of people
  7. I get more nervous performing a repertory 7  9 12  5
piece rather than improvising in front of
people
  8. I would prefer to improvise when I am 1  2 19 11
alone rather than in front of people
  9. The skill set helped me improvise 0  0 23  1
10. I feel that the skill set would help other 0  0 24  8
students improvise

Each subject stated his or her age, grade level, and years of study. Eighteen subjects studied
1–2 years, equally divided between treatment groups E1, E2, and control group E3, six in each.
Four elementary, four middle, and five high school subjects had studied 1–2 years. Three
elementary, six middle, and two high school subjects had studied 3–4 years. Two middle school
and five high school subjects had studied 5+ years. Overall, for those who had studied 1–2
years, it was found that members of the treatment groups exhibited lower levels of anxiety than
those of the control group. Furthermore, members of the treatment groups who had studied
1–2 years had lower levels of anxiety following a musical performance in front of the public.
Those who had studied 3–4 years exhibited higher levels of anxiety.
Results from the Pearson Product-Moment Correlation Coefficient found that there was a
weak positive relationship between the numbers of years studied and levels of anxiety based on
the MARS scores for subjects who had studied 1–2 years. Although positive, findings indicated
that there was no relationship between the numbers of years studied and levels of anxiety
based on the MARS scores for subjects who had studied 3–4 or 5+ years.
Although means distributions showed that all subjects experienced some variation in
individual levels of anxiety regardless of their years of study, results were not significant. Based
on results from the Pearson Product-Moment Correlation Coefficient and regression analysis,
an association between the number of years that subjects studied and levels of anxiety based on
MARS scores was not significant.
In order to test the significance of subjects’ preference for improvising alone rather than in
front of an audience, treatment groups E1 and E2 were combined into one test group. All 24
subjects from the treatment groups and nine subjects in the control group responded to
statement 8 on the MARS (Table 3). A two-population means test with pooled variance was
used to determine whether any significant differences existed between treatment and control
groups’ preference for improvising alone rather than in front of an audience. Statement 8 was
a forced multiple choice question (no neutral answer). Results found no significant difference
(p > .05, p = .28) between the combined treatment group and control group preferences for
performing free improvisations alone versus in public before an audience.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 83

A chi-square test was applied to the Musical Anxiety Report Scale (MARS) comparing
responses of the control group with those of each treatment group to statements 2–8. Results
from the chi-square test related to statements 3 and 6 for treatment group E2 were significant.
Results for all other statements were not significant. Results related to statements 2–8 for
treatment group E1 were not significant. Based on these findings, subjects from both treatment
groups were more nervous performing a repertory piece rather than improvising in front of
people. Although the results supported the notion that free improvisation may reduce levels of
anxiety during a performance in front of an audience, findings were not significant regarding
subjects’ preference for improvising alone rather than in front of people.
Results from paired t-test for paired observations of male and female MARS scores were not
significant (p > .05). For male and female subjects who did not receive the treatment (E3), results
were not significant (p > .05). Findings showed that females from the treatment groups exhibited
lower levels of anxiety than males during a musical performance in front of the public and
without an audience. On the other hand, females from the control group exhibited higher levels
of anxiety than males during a musical performance both in front of the public and without an
audience. Overall, both male and female subjects who received the treatment had lower levels
of anxiety whether performing with or without an audience.
The use of qualitative instruments such as subject interviews offered the reader insight into
attitudes and beliefs of students regarding the use of free improvisation as a treatment for
reducing anxiety while performing in front of an audience. Furthermore, it was believed that
this data may afford researchers a basis for further investigation into the social, psychological,
and cultural aspects of the family environment and its effect on the attitudes and beliefs that
individuals have regarding performance anxiety and its symptoms. The results were amplified
and triangulated against other data.
The following comments were derived from audio-recorded transcriptions of individual
sessions. They provided evidence in regard to subjects’ feelings towards free improvisation and
support the notion that free improvisation has a positive effect related to individual performance
anxiety levels. Subject comments were randomly selected.
Some subjects were hesitant about the concepts of improvising freely without having written
music as a guide. For example, Mary, middle school student, was uncomfortable about the
whole idea of playing the piano without music. She had no comment when asked how she felt
about free improvisation at the end of sessions 1 and 2. At the conclusion of session 3, she
explained that she liked the “middle part (event 2 – intervals),” but appreciated the “lowest part
(event 3 – rhythm)” the most. At the conclusion of session 6, the researcher asked: “What
events do you like the most?”. She replied, “Everything!”. Mary played a free improvisation and
a repertory piece during the performance.
Throughout her previous studies, Hanna, high school student, had difficulty reading music.
She had exhibited an unwillingness to practice the assigned music on a daily basis. She wel-
comed the concepts of free improvisation, agreeing to practice daily. After session 1, Hanna
remarked that she liked free improvisation. The researcher asked, “What do you like about it?”.
Hanna replied, “I can do my own thing.” In her final comments at the conclusion of session 6,
Hanna expressed her personal feelings about free improvisation:

You know what I just figured out, that whatever you play, not the third part (event 3), but in the other
part (event 2), you can sort of dream about ice skating and I try to match it up with what I play . . .
something like imagery.

Hanna played a free improvisation during the performance.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

84 Psychology of Music 41(1)

Middle school student, Martha, was concerned about hitting the “right” notes. Once Martha
realized that the “right” notes were hers to choose, she was willing to try free improvisation.
After session 1, Martha remarked that she “felt that there were sounds that did not mix.” When
asked what constituted the “right” sound, Martha replied, “I’ve trained my mind so that if
something doesn’t sound right when I’m doing it, I have to try something else to make it sound
right.” She added, “It’s hard when they don’t sound right.” When asked to explain what she
meant by “sound right,” Martha replied, “This sounds good (subject played the interval of a
major third), but this doesn’t (subject played the interval of a minor second several times).” At
the conclusion of session 6, Martha stated that free improvisation was “fun and easy.” Martha
played a free improvisation and a repertory piece during the performance.
Anna, elementary student, was anxious to begin learning the concepts related to free
improvisation. When asked how she felt about free improvisation, Anna giggled. Throughout
the treatment, she emphasized how much she liked free improvisation. During session 4 the
researcher asked, “In what way is free improvisation fun?”. Anna replied that it was a way to
“express myself.” At the conclusion of session 6, the researcher asked Anna to discuss her views
of free improvisation. She explained that it was a way of “making up your own thing that no
one else can make up.” Anna played a free improvisation during the performance.
During elementary student Joey’s session, the researcher asked, “How do you feel about free
improvisation?”. Joey replied, “I can make things happen, go wild!”. The researcher then asked,
“What was fun about free improvisation?”. Joey replied, “Well, instead of the book, I get to make
my own stuff up, and maybe one day it will be heard everywhere.” During session 3, Joey stated
that he liked free improvisation because he “did not have to read notes.” At the conclusion of
session 5, the researcher asked, “What do you think about free improvisation?”. Joey replied,
“It’s fun . . . because it’s fun making up your own songs instead of playing regular songs.” Joey
played a free improvisation during the performance.
During Nick’s first session, the researcher asked, “How do you feel about free improvisation?”.
Nick, middle school student, replied: “It tests you to see what you actually can do . . . if you really
know the piano.” The researcher asked, “How do you feel about the creative process?”. Nick
replied, “It was fun, fun, fun, fun.” When asked the same question at the conclusion of session 2,
Nick remarked, “It allows a person to play by ear . . . make more of something . . . have a melody
that no one would usually think of.” During session 3, the researcher asked, “What do you think
of event 3?” Nick replied, “Boom, boom, boom . . . it’s really active . . . it’s almost like a climax!”.
Nick discussed the similarities of free improvisation and event 3 to the music of the silent movie
era (early 20th century). His comments centered on how musicians would freely improvise
during the silent movie in order to provide musical imagery and enhance the viewing and
listening experience of the patrons: “For the old movies and stuff, since they had no sound . . . a
person had seen it lots of times . . . had to learn it . . . they would improvise . . . like the climax.”
At the conclusion of session 6, Nick stated that he felt that free improvisation was “awesome.” In
addition, when asked which event he preferred, Nick responded that he liked event 4 the most:
“because you can do anything.” Nick played a free improvisation during the performance.
During session 1, David, high school student, expressed his views regarding free improvisation:
“It was interesting . . . it makes me nervous not reading music . . . it’s kind of different . . . yet,
it’s a thing I’d like doing when I get use to it.” He responded to the same question at the
conclusion of session 2: “It seems like an easy way to play.” The researcher asked: “How do you
feel about free improvisation?”. David replied that he felt “good” about it (session 4). The
researcher asked, “Why?”. David replied, “I feel like I don’t have any responsibility in this piece.”
Responding to the same question during session 5, David replied, “I like it . . . I get to do what I

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 85

want.” The researcher asked, “How does it [free improvisation] make you feel?”. David replied,
“I try to think what would sound good . . . there’s no pressure or anything . . . even if it doesn’t
sound good, I can move on . . . it’s easy . . . you’re making it up, it’s your thing, but if you were
to play some else’s [music], you would have to try to play it like them.” The researcher asked,
“Which event appealed to you the most?”. David replied, “I like the rhythmic one (event 3) and I
kind of like the first one (event 1), up here, soft.” David played a free improvisation during the
performance.
The parent questionnaire served only to provide general information related to the attitudes
of parents with regard to free improvisation, performance anxiety, and the nature of the family
environment. Responses from the parent questionnaire found that parents generally felt that
music played an important role in family life and had a positive influence on their child. Most
parents (97%) agreed that music was important regarding its entertainment value,
overshadowing the importance of music as a means for expression, socialization, education,
relaxation, and enjoyment. In addition, all parents viewed music as a way to promote creativity,
self-confidence, self-esteem, discipline, and for developing motor skills. More than one-third of
the parents (38%) felt that music training was an indispensable part of mind development and
self-expression. Furthermore, the majority (78%) of parents agreed that musical activities
helped develop a child’s creativity, concentration, and problem solving. Here follows abbreviated
results of the questionnaire: (1) 13 nationalities or cultures; (2) all of the respondents had
family members who played a variety of instruments; (3) parents were evenly divided in their
views regarding their child’s practice habits and attitudes towards practice; (4) ninety percent
of parents were frustrated with the inability of their child to approach practicing with a positive
attitude; (5) 15 parents (54%) felt that free improvisation was a wonderful creative outlet,
would encourage musical growth, and was an important vehicle from which children could
express their feelings; (6) although two-thirds of parents stated that their child was reluctant to
perform in front of others, only 13 (45%) felt that their child had exhibited “stage fright”; and
(7) 28 parents (97%) felt that performing music was beneficial to their child.
Results of the study found that (1) significant differences in terms of anxiety levels were
found between treatment and control groups; (2) significant differences were found between
the levels of students’ state anxiety prior to and following the public performance of a free
improvisation; and (3) significant differences were found between students’ individual anxiety
levels during their public performance of a free improvisation and repertory piece. Furthermore,
no significant differences were found in the levels of students’ state anxiety prior to or following
the public performance of repertory pieces nor were significant differences found between
treatment and control groups’ preference for improvising alone rather than in front of an
audience.

Discussion and concluding comments

Results from this research supported the notions that (1) total freedom of expression reduced
anxiety by allowing students to perform what they conceived “in the moment,” and that (2)
subjects performed with less anxiety when the audience was perceived to be without the ability
to judge the performances based on predetermined outcomes. Scores from the Musical Anxiety
Report Scale provided evidence that subjects were more nervous performing a repertory piece
rather than performing a free improvisation in front of people. All subjects who received the
treatment agreed that the skill set was helpful for developing a free improvisation and would be

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

86 Psychology of Music 41(1)

useful to others. Overall, results from the STAIC and MARS confirmed that the use of free
improvisation based on the skill set was valuable in reducing S-anxiety during the performances.
Comparisons of performance times showed that the average length of repertory pieces was
constant, while free improvisations during the performances were shorter than when performed
without an audience. Free improvisations lasted as long as five minutes when subjects were left
alone during treatments based on the evidence from audio recordings of individual sessions.
This offered further evidence that subjects were more comfortable performing alone.
Responses to the parent questionnaire regarding “stage fright” were of particular interest;
specifically, of the 20 parents who felt that their child was reluctant to perform in front of
others, only 13 stated that their child had exhibited “stage fright.” Although inconclusive,
these findings suggested that parents may have been reluctant to admit that their child exhibited
symptoms of performance anxiety for various reasons: (1) embarrassment caused by their
child’s resistance to go on stage; (2) unawareness of the symptoms related to anxiety; and (3)
not having solutions for reducing anxiety prior to or during a public performance. In addition,
responses from the questionnaire and comments from interviews found that parents, in
general, were aware that: (1) “stage fright” (performance anxiety) is a problem with most
music students; (2) children needed to be in a comfortable setting – in front of family, friends,
or alone – in order to perform with confidence; (3) students had a desire to enjoy playing music
without having to worry about playing wrong notes; (4) children enjoyed creating music; and
(5) children enjoyed free improvisation as a way to engage their imagination.
Only one type of musical instrument (piano) was used in this study, which limited the
participation of music students who may have had an interest in free improvisation, but who
did not play the piano nor met the other requirements for participation in this study.
Generalizations related to the applications of the skill set to other instruments may be limited,
even under similar conditions. In addition, this study examined a particular approach to free
improvisation pedagogy, using techniques and repertory exclusive to the piano that may not be
equivalent to other kinds of free improvisation instruction.
The use of free improvisation rather than pharmacological intervention, psychological
analysis, and other experimental treatments as a means for reducing anxiety during a public
performance may encourage individual musicians to explore different aspects of musical
creativity as a way to overcome stress, nervousness, and other symptoms that are manifestations
of anxiety. Free improvisation may also provide musicians the means to perform without being
adversely affected by the presence of an audience. Furthermore, free improvisation may provide
a balanced approach, giving students and teachers alike an opportunity to discover or rediscover
musical creativity. Results from this study supported the notion that, by applying principles of
free improvisation based on the skill set, students were able to find their own individual levels of
creativity and, in so doing, reduce their individual levels of performance anxiety.
In order to gather supporting evidence, first a larger sample of music students who play
instruments other than piano may provide an opportunity for more conclusive results. Second,
a controlled laboratory environment would enable researchers to measure other variables
relating to physical as well as psychological conditions of the participants. Third, the gathering
of demographic variables would improve an understanding of social, cultural, and environmental
influences on performance anxiety. Although findings from the questionnaire and interviews
offer insight into the social, psychological and cultural aspects of the family environment, future
research is required to evaluate the relevance of these phenomena to performance anxiety and
its symptoms: social – the relationship between parents and children and their interactions;
psychological – the actions, traits, attitudes, thoughts, and mental states of individuals; and

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

Allen 87

cultural – how the ideas, customs, skills, and arts of a people or group influence the family
environment. Fourth, anxiety may be observed and measured during future performances in
order to evaluate the effectiveness of this treatment.
Although the STAIC Trait Anxiety Inventory was used to measure specific anxiety such as
performance anxiety, this particular music intervention was designed to relieve anxiety
pertaining to music performance situations, not general anxiety in musicians’ daily lives such
as career stress. Therefore, the STAIC Trait Anxiety Inventory might not be an appropriate
measure for performance anxiety. Previous research has suggested that the relationship
between the Trait Anxiety Inventory and State Anxiety Inventory can be explored in future
research (Kim, 2005). A study of this nature may find correlations between high scores of the
Trait Anxiety Inventory and high scores of the State Anxiety Inventory. This would enable
researchers to examine the notion that musicians with anxious personalities experience high
levels of performance anxiety. Though this study found that such an outcome is logical,
further research is suggested. It would be useful to know if an intervention can reduce levels
of high state anxiety for persons based on their personality structure (Kim, 2005).
Although outside the scope of this study, further research would help in establishing
significant relationships between anxiety and the effects of the home environment: the
influence of parental attitudes, biases, and prejudices on students. Furthermore, evidence may
support the notion that, without the inclusion of free improvisation during personal practice
time, anxiety may continue to adversely affect musical performances, causing musicians to
seek solutions for performance anxiety through pharmacological intervention, psychological
analysis, and other experimental treatments. Finally, comments from subject interviews
supported evidence that the particular skill set used in this study may provide a practical
approach for developing a treatment for musicians who suffer from performance anxiety.
In conclusion, this research demonstrated that free improvisation, as an individual
treatment, was beneficial for reducing anxiety in musical performance. Further investigations
may establish the use of this treatment as an alternative to pharmacological interventions used
in previous studies. Furthermore, free improvisation may be helpful as an enhancement for
psychological wellness in musical performance.

References
Allen, R. (1996). Success: Process or product. Percussive Notes, 34(2), 55–56.
Allen, R. (2007). Performance anxiety among elementary, middle school and high school piano students: A pilot
study. (Music education research study). Boston, MA: Boston University.
Allen, R. (2010). Free improvisation and performance anxiety among piano students. Doctoral dissertation,
Boston University, USA.
Appel, S. S. (1976). Modifying solo performance anxiety in adult pianists. Journal of Music Therapy, 13,
2–16.
Campbell, P. S. (2009). Learning to improvise music, improvising to learn music. In G. Solis & B. Nettl
(Eds.), Musical improvisation: Art, education, and society (pp. 119–142). Chicago, IL: University of Illinois
Press.
Darling, D. (2008). Return to child: Music for people’s guide to improvising music and authentic group leadership.
Goshen, CT: Music for the People.
Furth, H. G. (1969). Piaget and knowledge: Theoretical foundations. Englewood Cliffs, NJ: Prentice-Hall.
Gordon, E. E. (1997). Learning sequences in music: Skills, content, and patterns: A music learning
theory. Chicago, IL: GIA Publications.
Hamann, D. L. (1985). The other side of stage fright. Music Educators Journal, 71(8), 26–28.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015

88 Psychology of Music 41(1)

Kim, Y. (2003). Effects of improvisation-assisted desensitization, and music-assisted progressive muscle

relaxation and imagery on ameliorating music performance anxiety of female college pianists. Doctoral
dissertation, University of Kansas, USA.
Kim, Y. (2005). Combined treatment of improvisation and desensitization to alleviate music performance
anxiety in female college pianists: A pilot study. Medical Problems of Performing Artists, 20, 17–24.
LeBlanc, A., Jin, Y. C., Obert, M., & Siivola, C. (1997). Effect of audience on music performance anxiety.
Journal of Research in Music Education, 45(3), 480–496.
Lehrer, P. M. (1987). A review of the approaches to the management of tension and stage fright in music
performance. Journal of Research in Music Education, 35(3), 143–153.
Orcher, L. T. (2005). Conducting research: Social and behavioral science methods. Glendale, CA: Pyrczak
Publishing.
Rack, J. E. (1995). Performance anxiety in student musicians. Doctoral dissertation, Pennsylvania State
University, USA.
Roland, D. J. (1993). The development and evaluation of a modified cognitive-behavioural treatment for musical
performance anxiety. Doctoral dissertation, University of Wollongong, Australia.
Ross, A. (2007). The rest is noise: Listening to the twentieth century. New York: Farrar, Straus and Giroux.
Ryan, C. A. (2000). A study of the differential responses of male and female children to musical performance
anxiety. Doctoral dissertation, McGill University, Canada.
Sawyer, R. K. (2000). Improvisation and the creative process: Dewey, Collingwood, and the aesthetics of
spontaneity. Journal of Aesthetics and Art Criticism, 58(2), 149–161.
Spielberger, C. D. (1973). State-trait anxiety inventory for children. Retrieved 10 August 2007 from
http://www.mindgarden.com/products/staisch.html
Steptoe, A. (1982). Performance anxiety: Recent developments in its analysis and management. The
Musical Times, 123(1674), 537–541.
Wilson, G., & Roland, D. (2002). Performance anxiety. In R. Parncutt & G. McPherson (Eds.), The science
and psychology of music performance: Creative strategies for teaching and learning (pp. 47–62). New York:
Oxford University Press.

Author Biography
Robert Allen, Jr. received the Bachelor of Music degree in music performance from the
University of Northern Colorado at Greeley (1981), Master of Music degree in music
performance from the University of North Carolina at Greensboro (1985), and the Doctor of
Musical Arts degree in music education from Boston University (2010). From 1989–1997
Robert taught percussion at Rocky Mountain College and general music at Yellowstone Baptist
College concurrently. While at Yellowstone Baptist College, Dr. Allen developed the curriculum
for the music program and taught courses in theory, music history, form and analysis, piano,
conducting, and directed the choir. Dr. Allen established himself as a performing artist during
the 1980s, and his success as a teacher and lecturer enabled him to become a music clinician
for the Ludwig/Musser Drum Company, 1991–2007. In 1993, he was featured in a PBS
Special entitled “The Piston Stroke.” His article “Success: Process or Product,” which discusses
the issue of performance anxiety, was published in 1996 in Percussion Notes, a publication of
the Percussion Arts Society. His recordings include All Bach and Marimba Romance for solo
marimba, as well as Spanish Romance; Reverence; and My Brother Made Me Do It for marimba
with orchestra. Currently he maintains a busy private teaching studio with over 70 students,
many of whom have competed at the regional and national levels in music performance.
Several of his students have also received awards and scholarships in order to further their
careers. His teaching methods have helped many students attain their goals of becoming
professional music educators and performers.

Downloaded from pom.sagepub.com at EMORY UNIV on April 19, 2015